Jiangsu Bank Audit System Upgrade – Big Data Cluster Analysis System

lEmpower Complex Audit Workloads: By harnessing GBase 8a MPP Cluster’s strengths in columnar storage, intelligent indexing, and massively parallel processing, it enables aggregation, mining, and analysis of massive audit-related data, efficiently executing complex audit models to proactively detect and predict risks, ensuring sound banking operations;

lBreak Through Scalability Barriers: Leverage GBase 8a MPP Cluster’s horizontal scaling capability to protect existing investments and seamlessly scale the system in the future to support more data, applications, and users;

lEnhance High Availability: GBase 8a MPP Cluster’s security group architecture delivers transparent high availability, keeping services online even in the event of server failures;

lResolve the Cost-Performance Dilemma: Running GBase 8a MPP Cluster on commodity x86 servers reduces hardware investment by 80% compared to the legacy system, while delivering significantly higher performance.

As part of the audit system transformation project at Bank of Jiangsu, the GBase 8a MPP Cluster is composed of 4 computing nodes and 1 loading node, forming 2 security groups with 2 computing nodes in each. Data within each security group is backed up mutually to create a highly available cluster. Serving as the data platform layer for the audit system, GBase 8a MPP Cluster consolidates relevant business data and runs various complex audit models, delivering performance over ten times faster than traditional databases. It provides Bank of Jiangsu with high-performance audit services based on massive business data, and further supports the audit system’s evolving requirements for broader audit scope, business foresight, and comprehensiveness in the big data environment.

With 5TB of existing data, traditional databases suffer from increasingly slow query and analysis performance under high concurrency and heavy workloads, making it imperative to boost responsiveness. Audit-related SQL queries are highly ad-hoc and complex, placing stringent demands on the underlying database; the system must quickly execute complex audit models to deliver timely audit services. In terms of scalability, legacy architectures offer very limited expansion headroom in both compute and storage, failing to meet the future scaling requirements of audit operations. From a cost perspective, the system needs high cost-effectiveness to prevent unsustainable maintenance and scaling in later stages due to prohibitive expenses.

Bank of Jiangsu ranks among the top city commercial banks in China by business scale. As the bank continues to grow and its services become increasingly complex, the volume of audit-related data is surging, making traditional databases and conventional symmetric multi-processing (SMP) architectures unable to meet performance requirements. To ensure more effective system support, upgrading the existing audit system built on traditional databases has become both imperative and urgent.